The Walls Are Closing In On Tesla

Musk has destroyed every path forward.

Photo by Prometheus 🔥 on Unsplash

Thanks to his combination of ignorance, grift and hubris, Elon Musk is the king of self-destruction. Just take Tesla’s weird door handles. Back in 2016, Musk personally pushed for almost all vehicle functions, including the door handles, to be controlled by electric buttons or touchscreens. His own engineers and executives warned that this is a huge safety risk, as if the battery pack fails, as it very well can in a crash, then the doors can’t be opened by the occupants or rescue services. They argued for traditional, fully mechanical door handles, but Musk vetoed them for purely aesthetic reasons. He even pushed for the mechanical override, meant to be used in such emergencies, to be hidden, which made it hard, or even impossible, to use in emergencies. As such, entrapment is a significant safety risk for occupants of Teslas. At least 15 deaths have been attributed to this self-imposed safety flaw so far. This problem is so bad that China is banning Tesla-style car door handles by the end of the year, and the US Congress is in the process of passing a bill to do the same. As such, Musk and Tesla now look like utterly inept pillocks, and Tesla will have to spend millions of dollars redesigning their door handles and retooling their production lines. Now, if Musk can cock up so dramatically, and be so arrogant, ignorant and authoritarian over such a simple thing, how can you expect him to take Tesla forward successfully? Well, you can’t, and it shows, because every way Musk is desperately trying to grow Tesla is flat out failing.

Prepare for a very long and detailed rant about how utterly f**ked Tesla is in 2026.

Electric Vehicle Failure

I mean, take what is meant to be Tesla’s core business, selling EVs. Things are not going well for Tesla here at all!

BYD has officially surpassed Tesla as the world’s largest maker of battery electric vehicles (BEVs). Tesla’s annual sales declined by 9% in 2025 compared to 2024. But BYD’s annual sales increased 27.9% in 2025 compared to 2024. That meant it sold a whopping 38% more BEVs than Tesla in 2025!

But it wasn’t just BYD outselling Tesla. In Europe, VW’s annual BEV sales increased by 78% in the first half of 2025 compared to H1 2024, and outsold Tesla, whose annual sales in Europe declined by 33% in H1 2025 compared to H1 2024. This doesn’t include VW’s other subsidiary brands, like Audi, Cupra, Škoda, and Porsche. If you include these brands in VW’s BEV sales, they outsold Tesla three to one! The data currently available for the entire of 2025 suggests Tesla’s European performance was even worse in the second half of 2025, with a 27.8% decline in total sales compared to 2024 and strong sales from its competitors.

This is more than just political backlash to Musk. All the brands and EV models currently trouncing Tesla’s sales have something in common. The brands have dramatically expanded their EV line-ups, offering a plethora of new and unique choices. But, the best-selling among these new choices, and these EVs that are out-selling Teslas, are cheaper, yet still capable, more mass-market focused EVs, like the Renault 5, Škoda Elroq and VW ID.3. What’s more, the brands outselling Tesla, such as BYD, VW and Renault, have all had wild success designing, and building their own brilliant EV batteries in-house to reduce costs, and increase efficiency.

Why is all that important?

Well, if you cast your mind back to the start of the decade, Tesla planned to grow in this exact same way. They would expand their line-up with a pickup truck and a smaller, cheaper mass-market model. And all these models would use Tesla’s in-house 4680 battery, which would bring costs way down. BYD, VW and Renault’s success shows that this was the right idea for huge growth. But Musk personally got in the way of that.

Like with the door handles, Musk’s ignorant authoritarian micromanaging of the Cybertruck turned it into a design nightmare, and one of the biggest flops in automotive history. Likewise, his ignorant determination to use the unproven and unreliable proprietary dry-coating tech he bought for the 4680 battery means that even to this day, the 4680 is expensive and unable to scale properly.

Then, yet again, Musk vetoed his own executive’s advice and cancelled the more affordable “Model 2” he had promised for years, explicitly so Tesla could pour more cash into Tesla’s dead-end AI ambitions.

In other words, Tesla’s destroyed market share and rapidly declining sales and revenue can be solely blamed on Musk.

FSD Failure

But, surely, all that extra cash funnelled into Tesla’s Full Self-Driving (FSD) must have paid off? Well, no. Not at all.

At the end of 2024, Musk claimed that by the end of 2025, Tesla would have launched a robotaxi service using FSD that requires no human supervision, and would be available to over 50% of the US population! That simply hasn’t happened because FSD is still wildly unsafe and unable to meet regulatory requirements. In fact, FSD tracker, a site where Tesla customers who bought FSD self-report FSD’s performance, has found that the latest version (V14.2.2) has a critical disengagement rate (how often the driver has to step in to stop the system from crashing the car or driving in a horrifically illegal manner) of every 670 miles on average. In other words, if FSD were used as a properly unsupervised taxi and covered the same daily distance as a normal taxi, it would crash on average once every 3 days. And this is with self-reported data from FSD customers, so it is likely highly biased towards Tesla and cherry-picked. So, no wonder it isn’t safe enough for an unsupervised robotaxi.

Why is FSD so bad? Well, it is because back in 2021, Musk yet again vetoed his own expert engineers and scrapped all the FSD sensors, forcing it to run solely on camera feeds. This removed all redundancy in the system, meaning the computer vision AI and self-driving AI had to achieve near 100% accuracy for FSD to even function vaguely safely in optimal conditions (read more here). Well, we have known for years now that AI has diminishing returns (read more here), which not only means that getting AIs this accurate is simply impossible, but also makes it insanely expensive to even get them even slightly accurate.

So, again, Musk’s hubris, ignorance and arrogance have derailed a core part of Tesla’s business. And, it isn’t like it has drained a potential future business. No, this is badly affecting Tesla now.

Musk once claimed that a Tesla would be an appreciating asset, as self-driving would become a more expensive option as it became more capable. This meant that a $45,000 Tesla bought in 2020 with the $8,000 FSD option would be worth more than $45,000 today, because while the car might only be worth $10,000, the FSD would be worth more than $35,000. This was also meant to be Tesla’s big moneymaker, as if they could sell millions of Teslas with a $35,000 FSD option, then it would dramatically increase their income. Indeed, Tesla did increase the cost of FSD for a while, with it costing $15,000 in 2022. But no one was buying it at that insane price!

In 2024, Tesla slashed the price of FSD in half to get more customers through the door. Instead of costing $199 per month to subscribe, it was $99, and instead of $15,000 to buy outright, it was just $8,000. Tesla even offered free FSD trials to hundreds of thousands, if not millions, of Tesla drivers. Yet, data analysis from YipitData found that in 2024, only 2% of customers who tried FSD on these free trials actually went on to pay for it. In other words, 98% of people who used FSD thought it wasn’t worth the massively discounted $99 monthly price tag. This shouldn’t be surprising, as at the time, Tesla’s own data showed that FSD customers were only using it 15% of the time, heavily suggesting they did not trust or enjoy the system.

In other words, Tesla heavily slashed the price of FSD, destroying the myth that FSD would generate huge income for Tesla, and yet it still didn’t attract a meaningful number of customers.

This problem has persisted for Tesla.

In a recent earnings call, Vaibhav Taneja, Tesla’s CFO, said that “Total paid FSD customer base is still small, around 12% of our current fleet.” That is seriously appalling.

All the data also suggests that this is going to get worse from here. Consumer surveys have repeatedly found that FSD is actively driving customers away. FSD customers are using the system a tiny amount (I estimate they are currently only using it 20% of the time), which heavily suggests a dire lack of trust. People simply do not want to buy or use FSD, and it shows. On that same earnings call, Vaibhav Taneja, Tesla’s CFO, admitted that FSD revenue was down compared to 2024. This all heavily suggests that FSD revenue will likely continue to fall. Keep in mind that Tesla is spending more and more each year to operate and develop FSD, meaning the losses are likely to grow dramatically.

Musk sacrificed Tesla’s sure-fire growth in the EV market to bet the company’s future growth on FSD. Yet FSD isn’t just not growing, but actively dragging Tesla down, because his ignorance and ego ruined the product.

But, it wasn’t like Musk was warned of this… Right?

Well, the executive who advised Musk to focus on affordable EVs, and not self-driving, warned him that even in the best scenario, FSD would not be a profitable venture for Tesla.

So, FSD’s horrific performance and its massive financial failings are both squarely Musk’s fault for ignoring his own experts, who have been proven totally correct.

Robotaxi Failure

But FSD is more than just customer cars, right? Tesla has finally launched its Robotaxi service! Surely, that is going well. Right?

Well, no. Not at all.

Firstly, the service Tesla launched in Austin isn’t a true robotaxi service. A safety driver ready to take over at any time is still required by law, because the system is not safe enough. What’s more, there are currently fewer than 40 of these supervised robotaxis roaming Austin, far, far less than Musk promised. What’s more, only between 5 and 10 of these taxis are operational at any one time, which suggests Tesla has a permit to operate far fewer vehicles than the fleet size suggests. Again, this is likely due to safety issues with FSD.

Tesla isn’t just struggling to expand this service in Austin because of safety issues, but also across the US. For example, Tesla wanted to launch a similar service in San Francisco, but was denied a permit. As such, they are launching a ‘supervised’ robotaxi there, with a driver required to be in the driver’s seat and in control at all times, which is basically the same as if an Uber driver paid for FSD.

Tesla recently conducted its first fully autonomous drive, with no human in the car, on public roads in Austin. However, this was on suspiciously quiet roads while being followed closely by a chaser car, suggesting this was likely a very easy validation test or a thinly veiled PR stunt. Either way, it doesn’t move the needle forward for Tesla at all.

As such, not only is Tesla’s revenue from robotaxis currently nonexistent, but their inability to expand services or ability to operate safely enough to reduce human oversight means it will be nonexistent for quite a while still.

Compare this to Waymo. They currently operate 2,000 fully autonomous robotaxis with no safety driver in five major US cities, and do about 450,000 rides per week. Thanks to a long track record of relatively reliable, safe and cooperative operations, they can now dramatically expand operations, with services in Tokyo, London, Detroit, Las Vegas and other cities coming soon. This should mean that they will hit a million rides per week sometime this year, which equates to roughly a billion dollars in annual revenue.

How have Waymo done this? Well, they designed a self-driving system the way those Tesla engineers wanted to design FSD. The system has many layers of redundancy and constrains the AI’s tasks, enabling far, far safer operation, and easier, cheaper development.

Now, is Waymo profitable? No. But they are a damn lot closer than Tesla’s robotaxis. Not only do they actually have decent revenue, but their cost per mile is lower, and will get much lower soon (read more here). All this means that Waymo’s robotaxis don’t just have the legal permission to grow dramatically, but also the financial room to do so. A far lower cost per mile means that Waymo can expand its services far more cheaply than Tesla can, and reach break-even far more easily.

Waymo shows that if Musk recognised when he should put down his ego and listen to his own brilliant engineers, then the Tesla self-driving dream would still be alive. But, he didn’t, and now it is dead in the water, and it is all Musk’s fault.

Battery Failure

The only area Tesla grew in 2025 was its energy storage. But that masks a more troubling truth. Musk has killed Tesla’s own enormous battery potential.

You see, the batteries Tesla developed for its EVs would also be ideal for renewable energy storage. By selling the same battery in both EV and energy storage guise, they can reduce its cost per kWh through greater economies of scale. What’s more, as this battery is largely in-house, Tesla could benefit from the cost efficiencies and higher profits of vertical integration. This way, Tesla could undercut competitors, dominate the EV and renewable energy storage markets, and wildly profit from both markets.

This plan was seemingly going well. The Cybertruck, Megapack 3 grid-level battery and the Powerwall 3 home battery were all tipped to use Tesla’s in-house 4680, and enormous contracts with key material suppliers suggested Tesla was poised for a huge expansion of its own annual 4680 battery production. Investors felt like the profit tsunami was about to hit!

But the Cybertruck was a total flop, and that significantly reduced demand for 4680 cells. On top of this, because Musk has insisted on using an unproven, highly unreliable proprietary dry-coating technology he bought in 2019, 4680 production has had sky-high failure rates, leading to very low yields, which have pushed the price per cell up significantly and limited production capacity. As such, the 4680 has failed to hit its price targets, forcing Tesla to use LFP cells from CATL for the Megapack 3 grid-level battery and the Powerwall 3 home battery.

And now, it seems Tesla is totally giving up on the 4680. L&F were meant to be the major supplier of crucial high-nickel cathode material for it, but they have just cut the value of this contract from $2.9 billion to just over $7,000. In other words, L&F has effectively cancelled this deal, suggesting Tesla is winding down its 4680, as no other supplier is slated to replace them.

Tesla’s battery advantage was meant to be its secret sauce, and it is what made it valuable once upon a time. It was meant to help them stay one step ahead of the competition, undercut rivals, sell way more EVs and dramatically increase profits. Tesla giving up on the 4680 like this is a far greater blow than it might seem, and it destroys its own growth narrative.

But the failure of the 4680 isn’t entirely Musk’s fault, right? Well, surprise! It is.

If the Cybertruck hadn’t been a total failure, then 4680 demand would be high enough to make it worth Tesla’s time. But the Cybertruck was a total failure because of its truly awful design, and I’m not just talking aesthetically, but practically too. We know that Musk basically micromanaged the development of the Cybertruck and ensured it was precisely how he wanted it. So the Cybertrucks flop, and the knock-on fatal lack of demand for the 4680 can be squarely placed at Musk’s feet.

But, truth be told, the 4680 was dead on arrival and would have failed even if the Cybertruck sold well.

You see, back in 2020, all the major EV battery manufacturers were going in a very different direction from Tesla and developing 800-volt prismatic LFP battery packs because it was the obvious choice! These packs used technology proven to work at mass-market scale to make a battery pack that was easy to produce, cheaper per kWh than the 4680 could ever be, whilst being more energy dense, safer, longer lasting and charging just as fast, with the potential to charge far faster in the near future. So, even if the 4680 had hit all its targets, these batteries would still be better for both EVs and renewable energy storage.

In fact, these kinds of batteries are so much cheaper and better than the 4680 that Tesla chose to use them (in a 400-volt configuration) over the 4680 in both the Model 3 and Y. In fact, Tesla has been using these batteries since 2021. That is why Tesla never moved its entire line up to 4680 cells as promised, as it made no financial sense. This is also why the 4680 was so dependent on the Cybertruck to increase its demand because it was too expensive to deploy into the rest of Tesla’s line up.

Many analysts, engineers, and suppliers openly warned Musk back in 2020, when he announced the 4680, that it was highly risky, based on unproven tech, and could be obsolete on arrival. Even the CEO of CATL, Tesla’s largest battery supplier, told Musk to his face that the 4680 won’t work and that he doesn’t understand battery chemistry at all. Musk has had every chance to course correct, but hasn’t.

All that potential has been wasted, and Tesla is playing catch-up. Indeed, some reports suggest Tesla is retooling much of its 4680 lines to produce these LFP cells. But that is too little, too late. Someone else has already taken Tesla’s potential.

BYD’s blade battery is one of these high-voltage, prismatic LFP cells we have been talking about, and it is everything the 4680 wishes it was. Its staggeringly low-cost and high performance have enabled BYD not only to undercut the competition, causing its sales to skyrocket and outshine Tesla, but also push its profitability above Tesla’s through vertical integration. Last year, BYD started selling these cells as grid-level battery packs to compete with Tesla, and they are a truly astonishing 53% cheaper per MWh than Tesla’s. As such, it is set to dominate the grid-level energy storage market in the coming years and rake in even more revenue for BYD.

Again, BYD’s success proves that Tesla had the right idea, and had the enormous potential to grow that we all thought it did. But Musk’s ignorance, pig-headedness, ego, and hubris killed it.

Robot Failure?

All ignorant, egotistical megalomaniacs eventually have to contend with reality exposing their idiotic nature. Musk is no exception. Over the past few years, his EV, self-driving, robotaxi, and battery failures have been so spectacular that they now make front-page news. So, Musk has turned to the same tactic as all ignorant, egotistical megalomaniacs do when the world can see their inadequacy: pivot to something else entirely.

That is what the Tesla Optimus humanoid robot is. A desperate attempt to bury the cutting criticism and horrific reality, and still appear as the Stark-esque leader he was once seen as.

Unlike with battery technology, Tesla has no inherent advantage here. They are not a leader in AI, or real-world AI systems (as FSD’s failure proves). They also have no experience in developing humanoid robotics, integrating them with AI, or deploying them successfully. They do have the ability and experience to manufacture at scale, but that isn’t the big hurdle for humanoid robotics. No, the big challenge is making them work at all, and Tesla has no advantage here.

In fact, arguably, there isn’t a growth opportunity here at all. I have covered this before, but humanoid robots are inherently inefficient, needlessly challenging to deploy and wildly unreliable. They are PR tech-demonstrators that use anthropomorphising tricks to pull us in; they are not, and arguably never will be, disruptive or widely useful to any industry. That isn’t just me saying that; it’s the guy Musk initially brought in to head up Tesla Bot development, Chris Walti.

The instant Musk claimed the Optimus robot was more than just a shite demonstration of Tesla’s AI abilities, the project failed. It doesn’t matter that after more than three years of development, it is nothing more than a VR puppeteered animatronic (a level of technology Disney had 20 years ago). Even if it had worked as promised, it would still be mostly useless to the potential customers. The potential simply isn’t there.

But the current influx of humanoid robots has nothing to do with actually disrupting industries. Instead, it is meant to give the illusion of innovation and potential disruption. They ‘feel’ like a threat to industry, and their anthropomorphising nature enhances that dramatically, making people perceive them, the companies which build them and those which use them, as valuable. It is the exact same playbook which has made AI stocks so valuable.

But, even if this manipulative PR is the end goal of Optimus, Tesla is still set to fail. After all, you need a robot capable of fooling people, and the Tesla bot doesn’t even do that. However, there is a robot that not only fools people but is entering mass production, is owned by, and will be commercially used by, one of Tesla’s main rivals.

When you think of humanoid robots, only one name comes to mind: Boston Dynamics. The outfit has been making bleeding-edge robotics for over 30 years. We all watched agog 8 years ago, when Boston Dynamics showed off its Atlas humanoid robot, not just moving like a human, but executing perfect parkour moves. The latest iteration of Atlas is far sleeker and even more capable. It moves so effortlessly, and pulls off complex tasks with such ease, that it can be hypnotic.

In other words, while Tesla has no advantage in the humanoid robot space, Boston Dynamics is the global leader.

In 2021, Boston Dynamics was bought by one of the few companies that directly rivals Tesla: Hyundai. Now, under the South Korean stewardship, Boston Dynamics is mass-producing a commercial version of its Atlas robot, with plans to produce 30,000 annually by 2028, and Hyundai is planning to deploy them on vehicle production lines that very year. But this is no slapdash deployment, as Atlas will be used in tasks with proven benefit, such as parts sequencing.

Sure, this robot will likely be more expensive than the Tesla Bot. But crucially, it will actually work. Sure, I, and many others, highly doubt whether it will actually be more cost-effective or better than other solutions, but at least it can pass the very low bar of actually working, unlike Optimus. Atlas is enough to sell the illusion; Tesla’s Optimus simply is not.

As such, rather than making Tesla look good, Tesla’s stupid robot will make Tesla look moronic by drawing comparisons to one of its biggest rivals, who is far, far ahead of them in the robotics game. So, not only is Optimus a failure from the get-go, but the illusion it is trying to sell will backfire and paint Tesla out as totally incompetent by drawing comparison to Atlas! That means that Optimus won’t increase revenue, and very likely won’t increase Tesla’s idiotic share valuation either.

Going Forward

Every path forward for Tesla has been actively dismantled by Musk’s boundless ignorance, hubris, arrogance, ego, and megalomania.

Tesla will not collapse in 2026. At least, I don’t think it will. But with every path forward being so brutally cut off, its slow death will become more rapid, and more obvious. Musk has run out of places to turn and run out of false promises to make. He needs to reflect, accept his shortcomings, seek out experienced advice, and change course. There is still time and the cash reserves for Tesla to expand its line up with cheaper mass market alternatives that can sell like hot cakes, develop a usable and safe self-driving system, build industry-leading in-house batteries, and even develop the specialised robotics that are making a difference to industry rather than these useless humanoid ones. But Musk lacks the capacity; he could never be that vulnerable, and considering he has actively surrounded himself with yes-men on every front and fired any dissenting voices from his inner circle, he has made it so that he never has to be that vulnerable.

Tesla is a zombie company. The clock is ticking, and the walls are closing in; a crushing end is evidently in sight. Yet, rather than do anything to save it, Musk is playing a very 1940s-flavoured make-believe. Tesla is, functionally speaking, dead, because that is the trajectory Musk has locked it into.

---

Thanks for reading! Don’t forget to check out my YouTube channel for more from me, or Subscribe. Oh, and don’t forget to hit the share button below to get the word out!

---

Sources: Autoexpress, Bloomberg, New Atlas, The Verge, Engadget, Tech Radar, Futurism, Invezz, Electrek, CITA, The Verge, Reuters, BI, Electrek, Eleport, BSC, FSD Tracker, Teslarati, BBC, FHWA, Futurism, NATA, E Marketer, Futurism, Clean Technica, Electrek, Hyundai, Will Lockett, Will Lockett, Will Lockett, Will Lockett, Will Lockett, Will Lockett, Will Lockett, Will Lockett